CO 2 Circuit Breaker Arc Model for EMTP Simulation of SLF Interrupting Performance

Transcription

1 CO Circuit Breaker Arc Model for EMTP Simulation of SLF Interrupting Performance K. Udagawa, T. Koshizuka, T. Uchii, T. Shinkai, H. Kawano Abstract-- This paper presents a CO circuit breaker arc model for an Electro Magnetic Transients Program (EMTP) simulation of SLF interrupting performance. One Cassie arc model and two Mayr arc models were serially connected. A large post arc current was measured. The simulated post arc current using the arc model agreed well with measurements. The temperature profile of the arc was calculated, and it was shown that the CO arc has an arc core and an outer zone of arc column. Using EMTP simulations, the most severe SLF condition of the CO circuit breaker was found to be 7% or 8%. Keywords: CO circuit breaker, Arc model, EMTP, SLF interruption. R I. INTRODUCTION esearch on substitute gases for SF 6 gas, which is used in switchgear, is currently underway. Using CO gas as an arc-extinguishing medium for circuit breakers has attracted the attention of researchers []. A flow of a large post arc current of several amperes has been obtained for a short line fault (SLF) interruption of a CO circuit breaker, suggesting that many aspects differ considerably from those of a SF 6 gas circuit breaker []. To evaluate the SLF interrupting performance of a SF 6 gas circuit breaker with arc model calculations, we developed serially-connected three arc models consisting of one Cassie arc model and two Mayr arc models [], []. Using the arc models, we obtained measured results that allow us to quantitatively evaluate the SLF interrupting performance of a SF 6 gas circuit breaker. In this paper, we applied the serially-connected three arc models to evaluate the interrupting performance of a CO circuit breaker. Consequently, using arc model calculations, we could reproduce aspects of arc voltages, an aspect of a large post arc current and the interrupting success or failure for the circuit breaker. The calculations showed that the K. Udagawa is with Power & Industrial Systems Research & Development Center, Toshiba Co., Kawasaki, JAPAN ( of corresponding author: keisuke.udagawa@toshiba.co.jp). T. Koshizuka is with Power & Industrial Systems Research & Development Center, Toshiba Co., Kawasaki, JAPAN T. Uchii is with Power & Industrial Systems Research & Development Center, Toshiba Co. Kawasaki, JAPAN T. Shinkai is with Power & Industrial Systems Research & Development Center, Toshiba Co., Kawasaki, JAPAN H. Kawano is with Power & Industrial Systems Research & Development Center, Toshiba Co., Kawasaki, JAPAN Paper submitted to the International Conference on Power Systems Transients (IPST) in Delft, the Netherlands June -7, parameters differed between the SF 6 gas circuit breaker arc model and the CO circuit breaker arc model. Accordingly, we clarified the differences in arc parameters between the CO circuit breaker and the SF 6 gas circuit breaker by calculating arc radius and temperature profile. Moreover, using the serially-connected three arc models, we examined the most severe SLF condition for the CO circuit breaker. II. SLF INTERRUPTING PERFORMANCE WITH A CO CIRCUIT BREAKER A. Post Arc Current Fig. shows the waveforms of the post arc current measured at. ka- Hz-9% SLF conditions for a 7 kv CO circuit breaker []. For the measurements, the short circuit current was kept constant at. ka-9% and the di/dt of the injected current from the synthetic test circuit was varied. The arcing time was set to be constant at ms. When the gradient of the current-zero (di/dt) was small, almost no post arc current flowed. As di/dt increased, post arc current increased. The post arc current increased most at di/dt=% (given the rated condition of %), indicating approximately amperes. Meanwhile, wave peak value of the post arc current occurred at approximately µs after the current-zero and the post arc current flowed for approximately µs. In this test, interruption of current failed at di/dt=%. The post arc current with a high wave peak value and a long duration, which is shown in Fig., was not measured for a SF 6 gas circuit breaker. Therefore, the high wave peak value and long duration may be considered to be major features of SLF interruption for a CO circuit breaker. Current was measured through Rogowski coils installed near the circuit breaker. The measured signals were digitized at a sampling frequency of MHz and a resolution of bits current (A) di/dt=98% 8% % %(Failure) Fig. Measured post arc current waveforms of CO gas circuit breaker

2 B. Arc voltage Fig. shows measured arc voltages. The arc voltages were measured with a voltage divider. While di/dt was small, the extinction peak increased to approximately kv. As di/dt increased, however, the extinction peak decreased. The extinction peak gradually declined to about kv at a di/dt of % and further decreased at a di/dt of %, thereby recording a failed interruption. Meanwhile, the extinction peak for a SF 6 gas circuit breaker also showed a similar trend of declining arc voltage values as the interrupting conditions became increasingly severe. It will be necessary in particular to analyze the arc voltage to determine whether the arc model used to evaluate the SLF interrupting performance of a SF 6 gas circuit breaker is applicable to a CO circuit breaker % di/dt=98% %(Failure) 8% Fig. Measured arc voltages of CO gas circuit breaker (relation to time) III. EVALUATION OF ARC MODEL USING SERIALLY CONNECTED ONE CASSIE AND TWO MAYR MODELS A. Application of Arc Model in CO Circuit Breaker To evaluate the SLF interrupting performance of a SF 6 gas circuit breaker we developed serially-connected three arc models []. They consist of one serially-connected Cassie arc model simulating a large-current area, a serially-connected Mayr arc model simulating the area near the extinction peak, and a serially-connected Mayr arc model simulating the area near the current-zero. It was revealed that (i) the arc voltage for a SF 6 gas circuit breaker is almost constant in the current domain above several thousand amperes regardless of interrupting conditions, (ii) current values at the extinction peak are almost constant, and (iii) the extinction peak value varies depending on interrupting conditions. Consequently, we set the serially-connected three arc models for a SF 6 gas circuit breaker as described below. a) The arc parameter was assumed for the Cassie arc model based on (i) above and was set to be constant; b) The arc time constant was assumed for the Mayr arc model simulating the area near the extinction peak based on (ii) above, and was set to be constant regardless of interrupting conditions. Arc power loss was assumed in view of the extinction peak value based on (iii) above, and was varied depending on the interrupting conditions; and, c) The arc time constant of the Mayr arc model simulating the area near the current-zero was set to % of that of the Mayr arc model simulating the area near the extinction peak. The arc power loss for the former was set to % of the latter. With these settings, it is not necessary to calculate complicated arc parameters. We applied the above arc models for a SF 6 gas circuit breaker to a CO circuit breaker. Fig. shows the relationship between the arc voltages of the CO circuit breaker in Fig. and current. The values on the horizontal axis are logarithms. The time in the figure is that from the vicinity of the wave peak value of the interrupting current to the area near the current-zero. Fig. reveals the following points. (iv) The arc voltage shows almost constant values even if di/dt varies in the large-current area above approximately, amperes; (v) The current value at the extinction peak is almost constant at approximately amperes; and, (vi) The extinction peak value varies if di/dt, or an interrupting condition, varies. When applying the serially-connected three arc models to a CO circuit breaker, the parameter for the Cassie arc model can be set to a constant value based on (iv) above. The arc time constant for the Mayr arc model simulating the area near the extinction peak can be assumed based on (v) above. The arc power loss can be assumed based on (vi) above. As a result, the concept of the serially-connected three arc models may seem applicable to the CO circuit breaker, although specific parameter values for the Cassie and the Mayr arc models and the ratios of the parameters for the Mayr arc model simulating the area near the current-zero are predicted to differ from the corresponding values of the SF 6 gas circuit breaker. di/dt=98% %(Failure) % 8%. Fig. Measured arc voltages of CO gas circuit breaker (relation to current) B. Calculation of Interrupting Success or Failure Fig. shows a comparison of the measured arc voltage waveforms for a CO circuit breaker and those obtained using calculations for arc models. Fig. presents variations in arc voltage against current before the zero point under the condition di/dt=% shown in Fig.. In the calculations, a. ka- Hz-9%SLF test circuit used for actual measurements was simulated with the EMTP, and the arc models were described using the Models function of the

3 EMTP. The calculations were performed for a total time of 8 ms from the wave peak point to the post-zero point of the current. The parameters used for the respective arc models were as follows: a) Cassie arc model: Arc time constant of. µs and arc voltage of. kv b) Mayr arc model (Simulation for area near extinction peak): Arc time constant of. µs and arc power loss of kw c) Mayr arc model (Simulation for area near current-zero): Arc time constant of.66 µs and arc power loss of kw In Fig., aspects of the arc voltages were reproduced well by calculations from the large-current area to the current-zero. In particular, the voltage decay of the arc voltage from the extinction peak to the current-zero agreed well between calculations and measurements. In the figure, the C waveform shows the arc voltage for the Cassie model, the M waveform shows the arc voltage for the Mayr model simulating the area near the extinction peak, and M waveform shows the arc voltage for the Mayr model simulating the area near the current-zero. The arc voltage of the circuit breaker in the calculations is the sum of these three arc voltages. Simulation Measurement. Fig. Simulated arc voltage waveform (di/dt=%) Fig. shows a comparison of calculations and measurements for the post arc current after the zero point with regard to the calculations in Fig.. The aspect and magnitude of the post arc current were reproduced well by calculations using the arc models. Above all, it seems that the aspect of increasing current was reproduced well for time periods after the current-zero to approximately. µs and to the subsequent wave peak value of the current. The aspect of post arc current after the zero point agreed well with measurements by matching the aspect of arc voltage before the current-zero with that obtained from measurements. Furthermore, although it is not shown in the figure, interrupting success or failure could be reproduced under all di/dt conditions in Fig.. Consequently, it is revealed that SLF interrupting success or failure can be evaluated using the serially-connected three arc models for a CO circuit breaker. As described previously, the arc time constant and the arc power loss values for the Mayr arc model simulating the area M M C near the current-zero were % and % of the corresponding arc time constant and the arc power loss values for the Mayr arc model simulating the area near the extinction peak. Although other ratios were also used in the calculations, the values above most appropriately reproduced the aspects of arc voltages, the aspect of post arc current and the interrupting success or failure. Fig. Simulated post arc current (di/dt=%) IV. CONSIDERING PARAMETERS OF ARC MODEL A. Cassie Arc Parameters for CO Circuit Breaker that Differ from those of SF 6 gas Circuit Breaker Although it became clear that interrupting success or failure can be evaluated using the serially-connected three arc models even for a CO circuit breaker, the arc model parameters differed from those used of the SF 6 gas circuit breaker. Fig. 6 shows a scheme that presents the arc voltages of the respective arc models and the total arc voltage for the circuit breaker as the sum of the three arc voltages. The arc time constants for the respective arc models, which were used for both SF 6 gas and CO circuit breakers, are also input in the figure. C, M, and M are same as Fig.. Comparing both types of circuit breaker, the difference in the arc time constant for the Cassie arc model was significantly large at. µs for the SF 6 gas circuit breaker against. µs for the CO circuit breaker. We examined the differences by focusing on changes in arc radius in the area near the current-zero., Arc voltage (kv) current (A) Current Measurement Simulation Total arc voltage C M M SF6: 6 :.μs.6μs.6μs (%.6) CO: :.μs.μs.66μs (%.) Fig.6 Arc model and arc time constant in SF 6 and CO arc simulation

4 B. Method of Deriving Temperature Profile in CO Arc To calculate changes in temperature profile and arc radius of a CO arc, we used the Law of Conservation of Energy, which is used for calculating the temperature profile of a wallstabilized arc [], []. The equation () below is a conservation energy equation where the Joule heating of the arc balances with energy loss from thermal conduction and forced convective flow in the axial direction. σe / 7 d dt V r + rκ ρ( h h r dr dr l R σe = σ r d dr rκ I G G R = dt dr πrσdr / 7 h () : Joule heating of the arc : Energy loss from thermal conduction Where r: Radius, T: Temperature, α: Electric conductivity, E: Electric field, K: Thermal conductivity, p: Density, h: Enthalpy, V : Flow velocity, l: Arc length The above equation contains three hypotheses. The arc is axisymmetric; Local thermal equilibrium is a premise and the current domain where energy loss via radiation can be neglected is covered; and, Gas flows at K toward the arc in the axial direction. To solve equation (), the physical properties values of CO gas in Reference [6] were used. In addition, the following three points were hypothecated and the Runge-Kutta Method was used for the relevant calculations. The container wall had a nozzle throat with a radius of mm. The temperature at the mm radius was K. Current was varied at A, A, A, and A. The energy loss from forced convection was always the same value. C. Temperature Profile in CO Arc ) = V r ρ ( h ) l R : Energy loss from forced convective flow Fig. 7 shows the calculated result of the temperature profile of a CO arc in the radial direction. As for CO gas, the domain with high electric conductivity includes an area where temperature is almost unchanged, despite the change in current. There is another area where temperature rises along with an increase in current at the outer zone closer to the center. The former is assumed to be the outer zone of arc column while the latter is assumed to be the arc core. The change in temperature was insignificant with regard to electric conductivity in the arc for the Cassie model, which leads to the hypothesis that the change in arc conductance depends on the change in the cross-sectional area of the arc core and not to the change in electric conductivity. Meanwhile, electric conductivity in the arc of the Mayr model was significant depending on temperature, which leads to the hypothesis that the change in arc conductance is controlled by the change in electric conductivity rather than the change in the cross-sectional area of the arc core. Temperature (K) A A A A Core part Outer zone of arc column Radius (mm) a) Radial temperature profile in CO arc Conductivity (S/m) 8 6 A A A A Radius (mm) b) Electric conductivity in CO arc Fig.7 Simulated radial temperature and conductivity profile in and CO arc Fig. 7 also shows the relationship between electric conductivity and temperature profile, and the following points can be deduced therefrom. a) In the area corresponding to the arc core, electric conductivity diminishes along with the change in current. b) At the outer zone of the arc column, electric conductivity is almost unchanged. c) Consequently, some models presenting the arc core and the outer zone of arc column would be necessary for the area near the current-zero of CO gas: one is the Mayr model that presents the arc core and the other is the Cassie model that presents the outer zone of arc column. d) c) above agrees with the facts that the arc time constant for the Cassie model is small and the impact of the Cassie model is significant up to the area near the current-zero with regard to the serially-connected three arc models. e) Meanwhile, it is said that only the arc core exists for the SF 6 gas circuit breaker near the area of the current-zero. Accordingly, the Cassie model presenting the outer zone

5 of arc column does not have a great impact at the area near the current-zero. The preceding description can also be explained by the large arc time constant of the Cassie model in Fig. 6. D. Arc Model to Decide Interrupting Success or Failure It has been clarified that whether interruption succeeds or fails for the SF 6 gas circuit breaker depends on the Mayr arc model simulating the area near the current-zero in calculations for the serially-connected three arc models [], []. This concept agrees with the fact that interrupting success or failure for the SF 6 gas circuit breaker is decided within several µs after the current-zero. Hereafter, we examine arc models that are associated with interrupting success or failure for the CO circuit breaker. Fig. 8 shows the waveforms of the arc voltage using serially-connected two arc models (Cassie model and Mayr model simulating the area near the extinction peak) after eliminating one Mayr arc model (M) simulating the area near the current-zero from the calculations in Fig.. In Fig. 8, the aspect of the arc voltage from the largecurrent area to the extinction peak is not so different from that in Fig.. However, regarding the time from the extinction peak to the current-zero, the difference between measurements and calculations increased compared to Fig., showing that the decay of the calculated waveform has become more rapid than the measured decay. Simulation M. Fig.8 Comparison between measured arc voltage and simulation using arc models, (di/dt=%) C Measurement Fig. 9 shows a comparison of the calculated waveforms between the post arc current in Fig. and the waveform using the serially-connected two arc models, for which the magnitude of the post arc current almost doubled and the duration of the post arc current more than doubled. However, it shows that interruption succeeded for the serially-connected two arc models. The following points can be deduced from the results above. a) SLF interrupting success or failure can be evaluated even for a CO circuit breaker based on the serially-connected three arc models: b) SLF interrupting success or failure for a CO circuit breaker using the serially-connected three arc models depends on the Mayr model simulating the area near the extinction peak: and, c) The Mayr arc model simulating the area near the currentzero only has the effect of limiting the magnitude of the post arc current. current (A) Arc Models Arc Models Fig.9 Comparison between measured arc voltage and simulation using arc models, (di/dt=%) V. MOST SEVER SLF CONDITON FOR CO CIRCUIT BREAKER It is said that the 9% condition is the most severe among all of the SLF conditions for a SF 6 gas circuit breaker. As shown in Fig. 6, the preceding paragraph can also be explained by the fact that, for the serially-connected three arc models, interrupting success or failure depends on the Mayr arc model simulating the area near the current-zero, at which the arc time constant is small at.6 µs. The arc time constant for the Mayr model simulating the area near the extinction peak as the determinant factor of interrupting success or failure for a CO circuit breaker is approximately µs, as described previously. It may therefore be considered that the most severe SLF condition for a CO circuit breaker differs from that of a SF 6 gas circuit breaker. We, therefore, examined the most severe SLF condition for a CO circuit breaker by calculating whether interruption succeeds or fails in the case of changing SLF conditions based on the serially-connected three arc models. TABLE shows values of interrupting current and transient recovery voltage (TRV) under 9%, 8%, and 7% SLF conditions, which were calculated for the IEC-67- high-voltage alternating-current circuit breakers [7]. Fig. shows samples of the calculated TRV waveforms under 9% (L9), 8% (L8), and 7% (L7) SLF conditions. In the calculations, the circuit breaker was considered to be an ideal switch with the following connections: a single-phase line of Ω surge impedance is connected to one terminal of the circuit breaker while the other terminal is connected to a 7 kv power supply through inductance and capacitance so that TRV frequency at the power supply side becomes. khz at the rated interrupting current of. ka. The length of the line was set to allow the interrupting current to meet L9, L8, and L7 conditions. Based on Table and Fig., compared to the L9 SLF condition, the interrupting current (di/dt) and the dv/dt of SLF- TRV decrease under L8 and L7 SLF conditions, whereas the wave peak value of SLF-TRV increases.

6 Rate Condition 7kV-.kA- Hz Rate short-circuit breaking current Short-line breaking current Voltage at the instant of current interruption Peak value of first peak of line side TRV TABLE I CURRENT AND VOLTAGE VALUE AT SLF INTERRUPTION Fig. SLF-TRV wave shapes at L9, L8 and L7 Next, Fig. shows the calculated results for interrupting success or failure in the case of setting the circuit breaker for the above circuit with serially-connected three arc models, given the aforementioned current and TRV conditions. The area near the interrupting current-zero is emphasized in the figure. In Fig., the L8 and L7 SLF conditions are also added. All the parameters for the serially-connected three arc models are deemed to be identical. - L7 (Failure) - Time (us/div) Fig. Simulated post arc current waveforms; L7-L9 L9 L8 L7 I SC (karms)... I L (karms) di/dt.6.9. u = U m*(-i L/I SC) (kv) u L* 9... (kv) Peak factor k=u L*/u Rise time t L (us).7.6. Rated-of-rise of u L*/t L (kv/us).7..7 line side TRV Peak value of U m= U r/ recovery voltage (=u )(kv) Transient Peak voltage u m= U m(+.i L/I SC) (kv) Voltage across cb (kv) L9 L8 Time μs/div L7 L7 L9 L8 L8 (Failure) In Fig., interruption failed under L8 and L7 conditions. Although a large post arc current was recorded under the L8 condition, interruption succeeded. Under the L7 condition, interruption succeeded with the least post arc current among these conditions. Based on these results, the 7-8% SLF conditions are more severe than the 9% condition in the case of an interrupting current under SLF conditions for a CO circuit breaker. VI. CONCLUSION We examined the possibility of applying serially-connected three arc models to the interrupting performance of a CO circuit breaker. Our conclusions are as follows: a) We applied the serially-connected three arc models, which were developed to evaluate the SLF interrupting performance of a SF 6 gas circuit breaker, to a CO circuit breaker. Consequently, it was shown that they can be used to evaluate the interrupting performance of a CO circuit breaker. b) The flow of a large post arc current was measured under the 9% SLF condition after current was interrupted. The aspect and magnitude of the post arc current were reproduced by calculations using the arc models. c) The arc parameters were estimated for Mayr arc model and Cassie arc model based on measuring arc voltage waveforms. More work can be done to extract the arc parameters from measurements using a numerical analysis, for example a fitting procedure in matlab. Moreover, we examined differences in the arc time constant for the Cassie arc model between the CO circuit breaker and the SF 6 gas circuit breaker based on the temperature profile of arc. The following conclusions were obtained. d) Radial temperature profile of the axisymmetric arc was calculated by resolving the relevant conservation energy equation where the Joule heating of the arc brings a balance with the energy loss from thermal conduction and forced convective flow in the axial direction. e) It was found that the CO arc has an arc core and an outer zone of arc column. f) It could be explained that, unlike the SF 6 gas circuit breaker, the Cassie arc model has a significant impact even in the area near the current-zero for the CO circuit breaker, with regard to the serially-connected three arc models. Furthermore, the following conclusions regarding the most severe SLF conditions were obtained by drawing on the determinant factor on whether interruption succeeds or fails for the CO circuit breaker. g) It was shown that 7-8% SLF conditions may be more severe than the 9% SLF condition for a CO circuit breaker, based on arc model calculations. h) It was shown that interrupting success or failure for a CO circuit breaker is decided at the area near the extinction

7 peak based on the arc model calculations. The fact that the most severe SLF condition, which was presented in c) above, is not 9% because the arc time constant for the Mayr model simulating the area near the extinction peak is long was explained VII. ACKNOWLEDGMENT We used the physical properties values of the CO gas, which were calculated by Yasunori Tanaka, Professor at Kanazawa University, for the temperature profile calculations of the arc radius in this paper. We thank him for his collaboration in the appropriation of data. VIII. APPENDIX Fig. shows an experimental setup for measuring current and arc voltage at SLF condition. Current was measured through Rogowski coils, and arc voltage was measured a voltage divider. These measured signals were digitized at a sampling frequency of MHz and a resolution of bits by converter device called Front End. After that, the signals were transmitted to control units through optical fibers. Rogowski Coil IX. REFFERENCES [] T.Uchii, H.Kawano, T.Nakamoto, H.Mizoguchi Fundamental Propertiesod CO Gas as an Arc Quenching Medium and Thermal Interruption Performance of Full-Scale GCB Model IEE-J Trans. B, s Vol., No., pp69-7 [] T.Koshizuka, T.Shinkai, K.Udagawa, H.Kawano Simulation of SLF Interrupting Performance for SF 6 Gas Circuit Breaker based on Serially Connected Arc Models IEE-J Trans. B, Vol.9,No.7,pp9-9 9 [] T.Koshizuka, T.Shinkai, K.Udagawa, H.Kawano Circuit Breaker Model using Serially Connected Arc models for EMTP Simulation Paper 8 International Conference on Power Systems Transients (IPST9) in Kyoto, Japan June -6, 9 [] Y.Kito, T.Matsumura A Numerical Model of Wall-Stabilized Arc Sustained in the Forced Axial Air Flow IEE-J Trans A, Vol. No. pp [] Y.Kito, T.Matsumura Energy Loss in the Wall-Stabilized Arc Column with Forced Axial Flow IEE-J Trans A, Vol. No.8 pp.- 98 [6] Y.Tanaka, N.Yamachi, S.Matsumoto, S.Kaneko, S.Okabe, M. Shibuya Thermodynamic and Transport Properties of CO, CO -O, and CO - H at Temperatures of -K at Pressure of.-mpa IEE-J Trans B, Vol.6 No. pp [7] International Electrotechnical Commission 67- High-voltage alternating-current circuit-breakers X. BIOGRAPHEIS Keisuke Udagawa was born on August 9, 98. He received his M.S. degree in Computer Science from Waseda University, Japan. He joined Toshiba Corporation in 6. He is presently a researcher of the High Power Technology Group, engaged in the study of interruption phenomena and developed interrupting chamber. Mr. Udagawa is a member of IEE of Japan, and IEEE Voltage Divider Circuit Breaker Tadashi Koshizuka was born on June 9, 96. He received his B.S. degree in 989 and M.S. degree in 99, both in electrical engineering from Tokyo Denki University, Japan. In 99, he joined Heavy Apparatus Engineering Laboratory of Toshiba Corporation, Kawasaki, Japan. Mr. Koshizuka is a member of IEE of Japan, and IEEE Front End Scale up Circuit Breaker Coaxial cable Resistance Grounding conductor Front End Inslator Toshiyuki Uchii was born on March 9th, 97. He received the B.S. degree in applied physics from Tokyo University of Science, Tokyo, Japan, in 99, and the M.S. degree in energy engineering and the Ph.D. degree from Nagoya University, Nagoya, Japan, in 997 and 6, respectively. In 997, he joined Toshiba Corporation, where he has been engaged in research on arc interruption phenomena and development of high-voltage gas circuit breakers. Dr. Uchii is a member of IEEE, IEE of Japan, and Cigre. He is presently a member of working group A. of Cigre. Takeshi Shinkai was born in Japan on March, 969. He received his B.S. and M.S. degrees in electrical engineering from Waseda University, Tokyo, in 989 and 99. In 99, he joined the High Power Laboratory, Toshiba corporation, and since then has been engaged in the development of gas circuit breker. Mr. Shinkai is a member of IEE of Japan and IEEE. Hiromichi Kawano was born in Oita Prefecture, Japan, on March 7, 96. He received his B.S. and M.S. degrees in electrical engineering from Kyoto University, Kyoto, Japan, in 98 and 98, respectively. Currently, he is a manager of the High Power Technology Group, Toshiba Corporation, Kawasaki where he joined in 98. He has been engaged in the development and design of gas-insulated switchgears and the study of interruption phenomena. Mr. Kawano is a member of the IEE of Japan Fig. Experimental setup for measuring current and arc voltage